def battle(max_size, programs):
permutations = itertools.permutations(programs)
pairs = set([])
triples = set([])
cumulative = {}
for p in programs:
cumulative[p] = 0
for p in permutations:
pairs.add(tuple(sorted(p[:2])))
if len(programs) > 2:
triples.add(tuple(sorted(p[:3])))
for mapfile in glob.glob('maps/*.json'):
with open(mapfile) as f:
themap = rwjson.readJson(f)
if len(themap['rivers']) > max_size:
print mapfile,'is too long with size', len(themap['rivers'])
continue
print '\ntesting map', mapfile, 'with size', len(themap['rivers'])
for pair in pairs:
scores = arena.arena(mapfile, pair)
ranks = rank_scores(scores)
print 'ranks are', ranks, 'and scores are', scores
for p in ranks:
cumulative[p] += ranks[p]
for pair in triples:
scores = arena.arena(mapfile, pair)
ranks = rank_scores(scores)
for p,v in ranks:
cumulative[p] += v
print 'cumulative so far:', cumulative
print 'cumulative score:', cumulative
def get_simulation_details():
"""Get simulation specifications from the user."""
data = {}
mode = input(">>>Select Mode<<<\n[1] Evolution\n[2] Learning\n")
mode = 'evolution' if mode == 1 else 'learning'
data['mode'] = mode
arena_no = input(">>>Select Arena<<<\n[1] Arena 1 (687x371)\n[2] Arena 2 \
(500x270)\n[3] Arena 3 (800x432)\n")
data['arena'] = arena_no
ar = arena.arena(arena_no)
model = input(">>>Select Neuronal Model<<<\n[1] Multi-timescale Adaptive \
Threshold (MAT)\n[2] Leaky Intergate and Fire (LIF)\n")
model = 'mat' if model == 1 else 'iaf'
data['model'] = model
init = input(">>>Select Input Pose<<<\n[1] Random\n[2] Fixed\n")
init = 'random' if init == 1 else 'fixed'
data['init'] = init
if init == 'fixed':
x = input(">>> Enter initial X-Position (max %d)<<<\n"""\
% ar.maximum_length())
y = input(">>>Enter initial Y-Position (max %d)<<<\n"""\
% ar.maximum_width())
theta = input(">>>Enter initial orientation angle<<<\n")
data['x_init'] = x
data['y_init'] = y
data['theta_init'] = theta
if mode == 'evolution':
# Evolution
pop_id = input(">>>Select Population id<<<\n[1] Population 1\n[2] \
Population 2\n[3] Population 3\n")
data['population'] = pop_id
generations = input(">>>Number of Generations<<<\n")
data['generations'] = generations
else:
#Learning
runs = input(">>>Number of Runs<<<\n")
data['runs'] = runs
time = input(">>>Duration of Simulation in Seconds<<<\n")
data['time'] = time*1000
interval = input(">>>Duration of Interval in Milliseconds<<<\n")
data['interval'] = interval
return data
def battle(max_size, programs, vis):
permutations = itertools.permutations(programs)
pairs = {}
for i in range(len(programs) + 1):
pairs[i] = set([])
cumulative = {}
games = {}
for p in programs:
cumulative[p] = 0
games[p] = 0
for p in permutations:
for l in range(2, len(programs) + 1):
pairs[l].add(tuple(p[:l]))
groups = []
for l in range(2, len(programs) + 1):
pairs[l] = list(pairs[l])
random.shuffle(pairs[l])
groups += pairs[l][:len(pairs[2])]
jobs = []
for mapfile in glob.glob('maps/*.json'):
with open(mapfile) as f:
themap = rwjson.readJson(f)
if len(themap['rivers']) > max_size:
print mapfile, 'is too long with size', len(themap['rivers'])
continue
for pair in groups:
jobs.append((mapfile, pair))
random.shuffle(jobs)
for mapfile, pair in jobs:
print ' {} {}'.format(mapfile, pair)
with open(mapfile) as f:
themap = rwjson.readJson(f)
scores = arena.arena(mapfile, pair, vis=vis)
ranks = rank_scores(scores)
for p in ranks:
cumulative[p] += ranks[p]
games[p] += 1
for (s, p) in scores:
print ' {} [{:.2f}]: {} ({})'.format(
ranks[p], cumulative[p] / games[p], p, s)
for p in programs:
print '{}: {}'.format(cumulative[p], p)
def do_arena(self, line):
_arena = arena(self.line_bot_api, self.event)
return _arena.process(line.parsed.args)
runs = input(">>>Number of Runs<<<\n")
data['runs'] = runs
time = input(">>>Duration of Simulation in Seconds<<<\n")
data['time'] = time*1000
interval = input(">>>Duration of Interval in Milliseconds<<<\n")
data['interval'] = interval
return data
if __name__ == '__main__':
data = get_simulation_details()
arena = arena.arena(data['arena'])
if data['mode'] == 'evolution':
# Evolution
population, num_individuals = ev.load_population(data['population'])
average_fitness, average_connectivity, best_of_generation = [], [], []
elite = [0, 0]
results.set_results_path(data['init'], data['model'], data['arena'],
data['population'])
print results.path
for gen in range(data['generations']):
# Create a separate folder for each generation results
results.create_generation_folder(gen)
print 'generation: %d' % (gen + 1)
generation_log = []
for i in range(num_individuals):
print i
